Mixed-effects model with lagged variable

[Naomi Baran]

Dear community,

I have a dataset of 115 completed observation rows (or event) of 75 different companies for the years 2004 to 2016. I have 6 independent variables and 3 control variables. As my dataset has a lot of missing values (each company has on average just 2 years of observation over the 13 years), I was advice to use the one-level mixed effect model.

However, one of my control variable is a lagged variable of the dependent variable. Does it work to include this lagged variable in my model using the mixed-effect? If yes, should I change something in my code and if not which model should I use?

Many thanks in advance for your answer.

Best,

Naomi

[Carlo Lazzaro]

Naomi:
welcome to ths forum.
I fail to get the relationship between missing values and intercept only -mixed- model.
Missing values remain what they are regardless the regression model.
Hence, you should first diagnose the mechanism underlying the missingness in your dataset and decide if and how you can deal with them.
That said, if you have a lagged dependent variable among predictors you should consider dynamic panel data regression models.